Mesh Adaptation and Shape Optimization on Unstructured Meshes

In this SBIR CRM proposes to implement the entropy adjoint method for solution adaptive mesh refinement into the Loci/CHEM unstructured flow solver. The scheme will initially be developed and tested for ideal gases and will then be extended to encompass mixtures of thermally perfect/calorically imperfect gases. This approach will use the current remeshing algorithm in Loci/CHEM which utilizes nonstandard general polyhedral elements. The main objective is to provide a robust mesh adaptation scheme that will improve simulation accuracy while reducing overall computational costs. The principal incentive to NASA is to make large-scale, complex flow simulations more accurate and affordable so that their benefits can be fully realized within the design cycle. During Phase I we will perform mesh adaptations for a number of geometries and flow conditions of interest to NASA. We will use these simulations to evaluate the robustness and effectiveness of the new adaptation scheme to improve accuracy, and reduce overall computational cost.During Phase II we will implement a full adjoint scheme into Loci/CHEM. This is a natural extension of Phase I, and will not only allow for complete output-based mesh refinement capability, but will also allow Loci/CHEM to be used for uncertainty estimation and as a shape/geometry optimization tool. The combination of solution adaptive mesh refinement, shape optimization, and uncertainty estimation will provide NASA with a high-confidence, predictive tool for development and assessment of innovative aerodynamic concepts over a wide range of flight regimes.